Management of computer networks and network elements has become increasingly more complex over the years just as the elements and networks have increased in complexity. It is not sufficient to state that a network element or link between elements is “up” as opposed to “down”. The performance of each network element and link between elements needs to be assessed in order to provide a complete view of the network. In order to assess performance, key performance indicators (KPIs) are used.
KPIs can include items such as “equipment temperature” to “traffic” and “dropped calls”. In wireless networks, it is very common for a single element to offer several thousand such KPIs. As many different network elements can constitute the complete network, the total domain of available KPIs can very quickly reach tens of thousands to hundreds of thousands KPIs, each collected per network element and for each time interval (for example, every 5 to 15 minutes).
Network performance management applications collect all these KPIs for each monitored network element on a periodical basis (for example, every hour or every 15 minutes). This collection can very quickly lead to over a billion KPIs to monitor, on a daily basis. The multiplication of KPIs can therefore mean that a very large computer will be needed in order to collect, store, manage and report on this amount of data. Furthermore, in composite application management, a central system often collects information from several, disparate and heterogeneous sources, processes this combined information as a whole, and generates alarms when faults are detected.
It is desirable to have data storage systems that are well equipped to efficiently store and retrieve large amounts of data.